Porosity estimation by semi-supervised learning with sparsely available labeled samples

Luiz Alberto Lima, Nico Görnitz, Luiz Eduardo Varella, Marley Vellasco, Klaus Muller, Shinichi Nakajima

Research output: Contribution to journalArticle

7 Citations (Scopus)

Abstract

This paper addresses the porosity estimation problem from seismic impedance volumes and porosity samples located in a small group of exploratory wells. Regression methods, trained on the impedance as inputs and the porosity as output labels, generally suffer from extremely expensive (and hence sparsely available) porosity samples. To optimally make use of the valuable porosity data, a semi-supervised machine learning method was proposed, Transductive Conditional Random Field Regression (TCRFR), showing good performance (Görnitz et al., 2017). TCRFR, however, still requires more labeled data than those usually available, which creates a gap when applying the method to the porosity estimation problem in realistic situations. In this paper, we aim to fill this gap by introducing two graph-based preprocessing techniques, which adapt the original TCRFR for extremely weakly supervised scenarios. Our new method outperforms the previous automatic estimation methods on synthetic data and provides a comparable result to the manual labored, time-consuming geostatistics approach on real data, proving its potential as a practical industrial tool.

Original languageEnglish
Pages (from-to)33-48
Number of pages16
JournalComputers and Geosciences
Volume106
DOIs
Publication statusPublished - 2017 Sep 1

Fingerprint

Supervised learning
Porosity
porosity
geostatistics
estimation method
Learning systems
supervised learning
Labels
well
method

Keywords

  • Conditional random fields
  • Facies classification
  • Latent variable
  • Porosity estimation
  • Ridge regression

ASJC Scopus subject areas

  • Information Systems
  • Computers in Earth Sciences

Cite this

Lima, L. A., Görnitz, N., Varella, L. E., Vellasco, M., Muller, K., & Nakajima, S. (2017). Porosity estimation by semi-supervised learning with sparsely available labeled samples. Computers and Geosciences, 106, 33-48. https://doi.org/10.1016/j.cageo.2017.05.004

Porosity estimation by semi-supervised learning with sparsely available labeled samples. / Lima, Luiz Alberto; Görnitz, Nico; Varella, Luiz Eduardo; Vellasco, Marley; Muller, Klaus; Nakajima, Shinichi.

In: Computers and Geosciences, Vol. 106, 01.09.2017, p. 33-48.

Research output: Contribution to journalArticle

Lima, LA, Görnitz, N, Varella, LE, Vellasco, M, Muller, K & Nakajima, S 2017, 'Porosity estimation by semi-supervised learning with sparsely available labeled samples', Computers and Geosciences, vol. 106, pp. 33-48. https://doi.org/10.1016/j.cageo.2017.05.004
Lima, Luiz Alberto ; Görnitz, Nico ; Varella, Luiz Eduardo ; Vellasco, Marley ; Muller, Klaus ; Nakajima, Shinichi. / Porosity estimation by semi-supervised learning with sparsely available labeled samples. In: Computers and Geosciences. 2017 ; Vol. 106. pp. 33-48.
@article{166c4690dc904d7a927b60aebd4f9b93,
title = "Porosity estimation by semi-supervised learning with sparsely available labeled samples",
abstract = "This paper addresses the porosity estimation problem from seismic impedance volumes and porosity samples located in a small group of exploratory wells. Regression methods, trained on the impedance as inputs and the porosity as output labels, generally suffer from extremely expensive (and hence sparsely available) porosity samples. To optimally make use of the valuable porosity data, a semi-supervised machine learning method was proposed, Transductive Conditional Random Field Regression (TCRFR), showing good performance (G{\"o}rnitz et al., 2017). TCRFR, however, still requires more labeled data than those usually available, which creates a gap when applying the method to the porosity estimation problem in realistic situations. In this paper, we aim to fill this gap by introducing two graph-based preprocessing techniques, which adapt the original TCRFR for extremely weakly supervised scenarios. Our new method outperforms the previous automatic estimation methods on synthetic data and provides a comparable result to the manual labored, time-consuming geostatistics approach on real data, proving its potential as a practical industrial tool.",
keywords = "Conditional random fields, Facies classification, Latent variable, Porosity estimation, Ridge regression",
author = "Lima, {Luiz Alberto} and Nico G{\"o}rnitz and Varella, {Luiz Eduardo} and Marley Vellasco and Klaus Muller and Shinichi Nakajima",
year = "2017",
month = "9",
day = "1",
doi = "10.1016/j.cageo.2017.05.004",
language = "English",
volume = "106",
pages = "33--48",
journal = "Computers and Geosciences",
issn = "0098-3004",
publisher = "Elsevier Limited",

}

TY - JOUR

T1 - Porosity estimation by semi-supervised learning with sparsely available labeled samples

AU - Lima, Luiz Alberto

AU - Görnitz, Nico

AU - Varella, Luiz Eduardo

AU - Vellasco, Marley

AU - Muller, Klaus

AU - Nakajima, Shinichi

PY - 2017/9/1

Y1 - 2017/9/1

N2 - This paper addresses the porosity estimation problem from seismic impedance volumes and porosity samples located in a small group of exploratory wells. Regression methods, trained on the impedance as inputs and the porosity as output labels, generally suffer from extremely expensive (and hence sparsely available) porosity samples. To optimally make use of the valuable porosity data, a semi-supervised machine learning method was proposed, Transductive Conditional Random Field Regression (TCRFR), showing good performance (Görnitz et al., 2017). TCRFR, however, still requires more labeled data than those usually available, which creates a gap when applying the method to the porosity estimation problem in realistic situations. In this paper, we aim to fill this gap by introducing two graph-based preprocessing techniques, which adapt the original TCRFR for extremely weakly supervised scenarios. Our new method outperforms the previous automatic estimation methods on synthetic data and provides a comparable result to the manual labored, time-consuming geostatistics approach on real data, proving its potential as a practical industrial tool.

AB - This paper addresses the porosity estimation problem from seismic impedance volumes and porosity samples located in a small group of exploratory wells. Regression methods, trained on the impedance as inputs and the porosity as output labels, generally suffer from extremely expensive (and hence sparsely available) porosity samples. To optimally make use of the valuable porosity data, a semi-supervised machine learning method was proposed, Transductive Conditional Random Field Regression (TCRFR), showing good performance (Görnitz et al., 2017). TCRFR, however, still requires more labeled data than those usually available, which creates a gap when applying the method to the porosity estimation problem in realistic situations. In this paper, we aim to fill this gap by introducing two graph-based preprocessing techniques, which adapt the original TCRFR for extremely weakly supervised scenarios. Our new method outperforms the previous automatic estimation methods on synthetic data and provides a comparable result to the manual labored, time-consuming geostatistics approach on real data, proving its potential as a practical industrial tool.

KW - Conditional random fields

KW - Facies classification

KW - Latent variable

KW - Porosity estimation

KW - Ridge regression

UR - http://www.scopus.com/inward/record.url?scp=85020020104&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85020020104&partnerID=8YFLogxK

U2 - 10.1016/j.cageo.2017.05.004

DO - 10.1016/j.cageo.2017.05.004

M3 - Article

AN - SCOPUS:85020020104

VL - 106

SP - 33

EP - 48

JO - Computers and Geosciences

JF - Computers and Geosciences

SN - 0098-3004

ER -